计算机科学 > 密码和安全
[2026年10月27日提交]
标题：通过预训中毒获得代理后门
查看 PDF HTML( 实验)
摘要：大型语言模型在从公共网络中刮出的庞大的，未经检验的蝎子身上预训。 这对在预训数据中植入有毒内容的攻击者造成了安全风险。 我们首次显示这种中毒可以在agent安装一个远程代码执行后门，导致它在用户的机器上执行攻击者选择的命令。 为了测量威胁结束，我们从零开始将Qwen3型号从0.6B,1.7B和4B的尺寸进行预训，然后利用清洁的管道将每辆后排车分别装入CLI agent. 用我们的新毒药设计将预训服的0.1%毒药植入后门：部署的agent在扳机出现时执行恶意命令，同时正常地在干净的输入上操作。 攻击成功率(ASR)随着模型尺度的增加而增加，4B模型以0.1%的毒率达到大约92%的ASR，即使中毒率下降到0.0005%，也保持阳性。 我们的代码可在此 https URL 中获取。
其他特征
当前浏览上下文：
(cs.CR) (中文(简体) ).
参考资料和引用
正在装入。..
书目和引用工具
书目浏览器( 什么是浏览器？)
连接文件( 什么是连接文件？)
Litmaps (什么是 Litmaps? ).
点击智能引用( 什么是智能引用？
与本条有关的代码、数据和媒体
阿尔法XIV (什么是AlphaXiv吗？)
文档的催化X代码查找器( 什么是催化X?)
DagsHub(什么是DagsHub?) (英语).
Gotit.pub(什么是GotitPub?) 互联网电影数据库(IMDb)上"GotitPub"?
Hugging Face(什么是Huggingface?
科学卡斯特(科学卡斯特是什么？
演示文稿
建议和搜索工具
影响花(什么是影响花？)
CORE推荐人(CORE是什么？)
arXivLabs：与社区合作者的实验项目
arXivLabs是一个允许合作者在我们的网站上直接开发并分享新的arXiv特性的框架。
与arXivLabs合作的个人和组织都拥护并接受了我们的开放、社区、卓越和用户数据隐私的价值观。 arXiv致力于这些价值观，只与坚持这些价值观的伙伴合作。
有个能给ArXiv社区带来价值的项目吗？学习更多关于arXivLabs。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Computer Science > Cryptography and Security
[Submitted on 27 Oct 2026]
Title:Agentic Backdoors via Pretraining Poisoning
View PDF HTML (experimental)
Abstract:Large language models are pretrained on vast, uncurated corpora scraped from the public web. This creates a security risk from attackers who plant poisoned content in the pretraining data. We show for the first time that such poisoning can plant a remote code execution backdoor in an agent, causing it to execute attacker-chosen commands on the user's machine. To measure the threat end to end, we pretrain Qwen3 models from scratch at sizes 0.6B, 1.7B, and 4B, and then post-train each into a CLI agent using a clean pipeline. Poisoning only 0.1% of the pretraining corpus with our new poison design implants the backdoor: the deployed agent executes the malicious command when the trigger is present while behaving normally on clean inputs. The attack success rate (ASR) increases with model scale, with the 4B model reaching roughly 92% ASR at a 0.1% poison rate and staying positive even when the poisoning rate drops to 0.0005%. Our code is available at this https URL.
Additional Features
Current browse context:
cs.CR
References & Citations
Loading...
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.